China's vision for global AI governance: From principles to practice

ByXiao Qian

In early July in Geneva, I attended a series of discussions during the AI for Good Global Summit 2026, the World Summit on the Information Society (WSIS), and several closed-door policy dialogues on AI governance. One topic repeatedly surfaced across conversations among governments, international organizations, industry, and academia: The world is entering an era in which AI development is accelerating while global governance is increasingly fragmented. If left unaddressed, this gap risks undermining not only technological progress but also international peace, development, and trust.

One message from the UN Independent International Scientific Panel on AI particularly caught my attention. The report argues that the emerging "AI divide" is no longer simply about who has access to technology. Increasingly, it is about who possesses the capacity to shape AI's future. This has created an increasingly unequal distribution of AI capabilities, making it more difficult for developing countries to participate meaningfully in the next wave of technological transformation. Without broader international cooperation, AI risks becoming another source of global inequality rather than an engine for shared prosperity.

A second challenge lies in the emergence of data and computing barriers. Data localization requirements, export controls on advanced chips, and restrictions on access to computing infrastructure are creating new digital divides. While every country has legitimate security concerns, excessive technological fragmentation may ultimately reduce innovation, increase costs, and make global governance even more difficult.

Third, the international governance landscape itself remains highly fragmented. In the meantime, advanced AI models increasingly demonstrate capabilities that cross traditional sectoral boundaries. However, different countries are developing AI regulations according to their own political systems, legal traditions, and economic priorities. These diverse approaches are understandable, yet without greater interoperability they could produce conflicting standards, duplicated compliance burdens, and regulatory uncertainty for both governments and industry. Global AI governance cannot succeed if every country builds an entirely separate rulebook.

Against this backdrop, China has gradually articulated a governance vision that seeks to move international discussions beyond competition toward broader cooperation. Rather than treating AI governance solely as a matter of risk control, China's approach emphasizes balancing innovation, security, and development. This philosophy can be seen across a series of recent initiatives.

The Global AI Governance Initiative, proposed by Chinese President Xi Jinping in 2023, provides the normative foundation. It emphasizes fairness, inclusiveness, international cooperation, and respect for different countries' development paths. More importantly, it recognizes that AI governance should not become an instrument for technological containment or geopolitical rivalry, but rather a platform for delivering global public goods.

The Global AI Governance Action Plan released in July 2025 moves beyond broad principles toward practical cooperation. It focuses on capacity building, technical exchanges, standards cooperation, talent development, and international dialogue. After all, governance succeeds not because countries issue declarations, but because they build institutions capable of implementing them.

Equally important is China's AI+ International Cooperation Initiative, which places development at the center of AI governance. AI should not be viewed solely through the lens of security. It also offers tremendous opportunities to improve healthcare, education, agriculture, disaster prevention, manufacturing, and sustainable development. For many countries in the Global South, these developmental applications are often more immediate than debates over frontier model governance. International cooperation should therefore focus not only on managing risks but also on expanding access to AI's benefits.

Taken together, these initiatives reveal a broader philosophy: Global AI governance should not force countries to choose between innovation and regulation, or between security and development. Effective governance requires all of them.

The international community needs opportunities to translate shared principles into practical cooperation. That is precisely why I believe the upcoming World Artificial Intelligence Conference (WAIC) in Shanghai deserves close international attention. Over the years, WAIC has evolved from a technology exhibition into one of the world's most important platforms where governments, international organizations, industry leaders, researchers, and civil society come together to discuss not only how AI should be developed, but also how it should be governed.

This year's conference is particularly significant. From discussions on the proposed World AI Cooperation Organization to international cooperation on AI safety and frontier model governance, AI capacity building for developing countries, AI for sustainable development, autonomous driving governance, and international standards, WAIC reflects China's efforts to move global AI governance beyond principles and toward practical action.

For me, WAIC has become one of the most interesting places to observe the evolution of global AI governance. It is where competing ideas meet, where policymakers engage directly with engineers, where international organizations interact with technology companies, and where proposals are increasingly transformed into concrete initiatives. At a time when many international discussions remain divided by geopolitical competition, platforms that encourage dialogue, experimentation, and practical cooperation are becoming more valuable than ever.

The future of AI governance will not be determined solely by who develops the most advanced models. It will also depend on who can build the broadest partnerships, the most effective institutions, and the greatest trust across borders. If the world hopes to ensure that AI serves humanity as a whole, then governance must move from principles to practice – and WAIC offers one place where that transition is already underway.

The author isVice Dean of the Institute of AI International Governance at Tsinghua University.